Accurate diagnosis for fault of switched reluctance motor power converter, on one hand, is depending on the information source of the fault, i.e., obtaining rich, authentic, and accurate fault information and symptoms is a prerequisite for the fault diagnosis; on the other hand, is depending on the fault diagnosing method. Therefore, to carry out fault diagnosis, not only rich, valid, and accurate fault information and data must be acquired, but also an advanced and efficient fault diagnosing method based on the limited fault data must be used, so as to accurately judge the fault type and accomplish fault diagnosis. Among existing fault diagnosing methods for the main circuit of switched reluctance motor switch converter, some utilize two fault characteristic quantities to identify a fault; while others utilize analog to digital conversion to extract fault characteristic quantities; all these methods have high requirements for hardware performance. An appropriate diagnosing method for fault of switched reluctance motor power converter, which is highly fault-targeted, can realize reliable and accurate fault diagnosis, and is highly practical, is an important guarantee for implementing fault-tolerant control of a switched reluctance motor system and for improving operating reliability of the switched reluctance motor system.